Sensor-based machine learning for workflow detection and as key to detect expert level in laparoscopic suturing and knot-tying

• Verfasst von: Kowalewski, Karl-Friedrich [VerfasserIn]
• Verfasst von: Garrow, Carly R. [VerfasserIn]
• Verfasst von: Schmidt, Mona Wanda [VerfasserIn]
• Verfasst von: Benner, Laura [VerfasserIn]
• Verfasst von: Müller, Beat P. [VerfasserIn]
• Verfasst von: Nickel, Felix [VerfasserIn]
• Titel: Sensor-based machine learning for workflow detection and as key to detect expert level in laparoscopic suturing and knot-tying
• Verf.angabe: Karl-Friedrich Kowalewski, Carly R. Garrow, Mona W. Schmidt, Laura Benner, Beat P. Müller-Stich, Felix Nickel
• E-Jahr: 2019
• Jahr: [November 2019]
• Umfang: 9 S.
• Illustrationen: 1 Diagramm
• Fussnoten: First online: 21 February 2019 ; Gesehen am 06.11.2019
• Titel Quelle: Enthalten in: Surgical endoscopy and other interventional techniques
• Ort Quelle: New York : Springer-Verlag, 2002
• Jahr Quelle: 2019
• Band/Heft Quelle: 33(2019), 11, Seite 3732-3740
• ISSN Quelle: 1432-2218
• DOI: doi:10.1007/s00464-019-06667-4
• Datenträger: Online-Ressource
• Sprache: eng
• Sach-SW: Artificial intelligence
• Sach-SW: Electromyography
• Sach-SW: Laparoscopic training
• Sach-SW: Laparoscopy
• Sach-SW: Machine learning
• Sach-SW: Myo armband
• Sach-SW: Neural networks
• Sach-SW: Skill assessment
• Sach-SW: Surgical education
• Sach-SW: Workflow analysis
• K10plus-PPN: 1681174995

Abstract

IntroductionThe most common way of assessing surgical performance is by expert raters to view a surgical task and rate a trainee’s performance. However, there is huge potential for automated skill assessment and workflow analysis using modern technology. The aim of the present study was to evaluate machine learning (ML) algorithms using the data of a Myo armband as a sensor device for skills level assessment and phase detection in laparoscopic training.Materials and methodsParticipants of three experience levels in laparoscopy performed a suturing and knot tying task on silicon models. Experts rated performance using Objective Structured Assessment of Surgical Skills (OSATS). Participants wore Myo armbands (Thalmic Labs™, Ontario, Canada) to record acceleration, angular velocity, orientation, and Euler orientation. ML algorithms (decision forest, neural networks, boosted decision tree) were compared for skill level assessment and phase detection.Results28 participants (8 beginner, 10 intermediate, 10 expert) were included, and 99 knots were available for analysis. A neural network regression model had the lowest mean absolute error in predicting OSATS score (3.7 ± 0.6 points, r2 = 0.03 ± 0.81; OSATS min.-max.: 4-37 points). An ensemble of binary-class neural networks yielded the highest accuracy in predicting skill level (beginners: 82.2% correctly identified, intermediate: 3.0%, experts: 79.5%) whereas standard statistical analysis failed to discriminate between skill levels. Phase detection on raw data showed the best results with a multi-class decision jungle (average 16% correctly identified), but improved to 43% average accuracy with two-class boosted decision trees after Dynamic time warping (DTW) application.ConclusionModern machine learning algorithms aid in interpreting complex surgical motion data, even when standard analysis fails. Dynamic time warping offers the potential to process and compare surgical motion data in order to allow automated surgical workflow detection. However, further research is needed to interpret and standardize available data and improve sensor accuracy.